Bug 89573

Summary:

[BYT] screen goes black and stays black with Linux 4.0-rc2 when X starts on Dell Venue 8 Pro 3845 baytrail tablet

Product:

DRI

Reporter:

Rene Wagner <rw>

Component:

DRM/Intel

Assignee:

Intel GFX Bugs mailing list <intel-gfx-bugs>

Status:

CLOSED DUPLICATE

QA Contact:

Intel GFX Bugs mailing list <intel-gfx-bugs>

Severity:

normal

Priority:

medium

CC:

intel-gfx-bugs, rw

Version:

unspecified

Hardware:

x86-64 (AMD64)

OS:

Linux (All)

Whiteboard:

i915 platform:

i915 features:

Attachments:

Description	Flags
kernel .config	none
dmesg output with default kernel command line	none
Xorg.0.log with default kernel command line	none
output of "xrandr --verbose" with default kernel command line	none
dmesg output with drm.debug=0x06	none
Xorg.0.log with drm.debug=0x06	none
output of "xrandr --verbose" with drm.debug=0x06	none
kernel log ("journalctl -k") with drm.debug=0x06 and xserver-xorg-video-intel-2.99.917	none
Xorg.0.log with drm.debug=0x06 and xserver-xorg-video-intel-2.99.917	none
output of "xrandr --verbose" with drm.debug=0x06 and xserver-xorg-video-intel-2.99.917	none

Description Rene Wagner 2015-03-13 20:52:19 UTC

Created attachment 114294 [details]
kernel .config

I'm trying to get Linux to work on a Dell Venue 8 Pro 3845 tablet that comes preinstalled with Windows 8.1. I'm running Debian Testing (Jessie) installed via Steve McIntyre's 64-bit kernel on 32-bit UEFI Debian Installer images [1]. If I go through the EFI boot menu, the device boots normally until X is started and the screen goes black, flashes briefly, and then stays black. If I prevent the intel X.org driver from loading by adding "nomodeset" to the kernel command line I can proceed to a working, albeit rather slow, Gnome session. This report is about getting accelerated graphics to work as well.

I noticed a number of Baytrail-related commits have gone into the kernel since 3.16.x which Jessie ships, so I've tried with a custom (yet unpatched) 4.0-rc2 build (and an unmodified Jessie user space).

The firmware is up to date (A02) and I should note that this is apparently not the same hardware as in earlier reports [2] which were for the model 3830 Venue 8 Pro.

I'm attaching the relevant configuration and log files. Xorg is started without an xorg.conf. There are two sets of logs one without and one with "drm.debug=0x06". The system crashed hard sometime after acquiring the first set.

The Xorg log does not contain errors. The kernel reports a "vblank wait timed out on crtc 0". One thing I noticed as somewhat odd is that xrandr lists the internal display (which AFAIK is connected via DSI) as "UNKNOWN1" rather than the "DSI1" which users of the 3830 model have reported but I don't know how significant that is. The internal display is the only display. There are no external connectors.

Let me know if I can provide any additional info or logs.

Regards,

Rene


[1] http://blog.einval.com/debian/CDs/Jessie-EFI_5.comments
[2] https://bugs.freedesktop.org/show_bug.cgi?id=82880

Comment 1 Rene Wagner 2015-03-13 20:53:33 UTC

Created attachment 114295 [details]
dmesg output with default kernel command line

Comment 2 Rene Wagner 2015-03-13 20:54:25 UTC

Created attachment 114296 [details]
Xorg.0.log with default kernel command line

Comment 3 Rene Wagner 2015-03-13 20:55:15 UTC

Created attachment 114297 [details]
output of "xrandr --verbose" with default kernel command line

Comment 4 Rene Wagner 2015-03-13 20:55:53 UTC

Created attachment 114298 [details]
dmesg output with drm.debug=0x06

Comment 5 Rene Wagner 2015-03-13 20:56:35 UTC

Created attachment 114299 [details]
Xorg.0.log with drm.debug=0x06

Comment 6 Rene Wagner 2015-03-13 20:57:31 UTC

Created attachment 114300 [details]
output of "xrandr --verbose" with drm.debug=0x06

Comment 7 Chris Wilson 2015-03-13 21:39:06 UTC

Start by installing xf86-video-intel-2.99.917 (xserver-xorg-video-intel) and update the Xorg.0.log and xrandr --verbose.

Comment 8 Rene Wagner 2015-03-13 22:04:54 UTC

Created attachment 114303 [details]
kernel log ("journalctl -k") with drm.debug=0x06 and xserver-xorg-video-intel-2.99.917

Comment 9 Rene Wagner 2015-03-13 22:06:05 UTC

Created attachment 114304 [details]
Xorg.0.log with drm.debug=0x06 and xserver-xorg-video-intel-2.99.917

Comment 10 Rene Wagner 2015-03-13 22:07:00 UTC

Created attachment 114305 [details]
output of "xrandr --verbose" with drm.debug=0x06 and xserver-xorg-video-intel-2.99.917

Comment 11 Rene Wagner 2015-03-13 22:14:55 UTC

Thanks for the suggestion, Chris. I've installed xserver-xorg-video-intel from experimental now and updated the log attachments. I've also updated the kernel log as the previous attachment had been truncated.

Unfortunately, there is no visible change in behavior. xrandr, however, now lists the display as DSI1.

# dpkg -l xserver-xorg-video-intel
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name                                        Version                    Architecture               Description
+++-===========================================-==========================-==========================-============================================================================================
ii  xserver-xorg-video-intel                    2:2.99.917-1~exp1          amd64                      X.Org X server -- Intel i8xx, i9xx display driver

Comment 12 Chris Wilson 2015-03-13 22:39:13 UTC

(In reply to Rene Wagner from comment #11)
> Thanks for the suggestion, Chris. I've installed xserver-xorg-video-intel
> from experimental now and updated the log attachments. I've also updated the
> kernel log as the previous attachment had been truncated.
> 
> Unfortunately, there is no visible change in behavior. xrandr, however, now
> lists the display as DSI1.

That and that the backlight is then detected was what I really wanted to check (in the ddx). I note that DSI *still* does not have the scaling modes exported.

Comment 13 Rene Wagner 2015-03-17 22:37:37 UTC

(In reply to Chris Wilson from comment #12)
> That and that the backlight is then detected was what I really wanted to
> check (in the ddx). I note that DSI *still* does not have the scaling modes
> exported.

Thanks again, Chris. Is there anything else I can do to help track this down?

Comment 14 Jesse Barnes 2015-04-02 18:11:13 UTC

Sounds like a dupe; we have some DSI work left to do.

*** This bug has been marked as a duplicate of bug 82880 ***

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.